| As an active ranging detection system, SAR can work day and night. It can work on both broadside and squinted modes. SAR can be employed on battleground scouting and used as fire control radar(FCR) for enhancing hit accuracy. It also can be used as a device of missile navigation. SAR imagery can be used on disaster detection, mine detection on civil aspect. A series of studys on squinted stipmap, spotlight, staring imagery are carried out. A detailed arrange is shown as follows:1. Current situation and history about SAR is discussed in this part. At the same time, the trend of development and research meaning is given.2. An algorithm combining SPECAN and azimuth non-linear scaling is proposed in this paper. SPECAN is used to solve azimuth Doppler spectrum alias in squinted spotlight mode. After the squinted case is equivalent to broadside case, range migration correction is implemented. The point target on the edge of image is defocused because of variance of Doppler frequency modulation rate in azimuth direction. The traditional low squint angle algorithm can not obtain a well focused SAR image. Doppler frequency modulation rate is approximate to one order variant, then the well focused region is enlarged. Finally, simulation is implemented for demonstrating the validity of the proposed method.3. The polar format algorithm is modified for squinted spotlight mode. In squinted spotlight case, azimuth spectrum is aliased. PFA can avoid spectrum alias because of dechirp implementation. In squinted case, PFA can cause azimuth spectrum non-uniformly spaced. A modified PFA is proposed to build a new imaging coordinate system. The simulation result is given in experiment. At the same time, the real data with 40 degree squinted angle is processed. The processed result is well focused which demonstrate the validity of the proposed method.4. A wide-swath squinted spotlight PFA algorithm is proposed for solving the problem that plane wave assumption causes small well-focused region. First, center of full aperture is selected as reference point. The full data is dechirped by this referent point. After azimuth coarsely focused, the whole image region is divided into several small regions. After dividing the whole region, the abandoned phase need to be compensated. After that, the center point of small region is selected as reference point for dechirping the small region data. The image resolution is not sacrificed because of Doppler spectrum is kept. After dechirp implementation, polar format algorithm is applied to the segmented image region. The deformed sub-image is obtained after two dimension sinc interpolation. The subimage resolution is approximate to the whole image resolution. Geometry deformation correction of sub-image is needed. Finally, all sub-images is combined into a whole image. This method can enlarge the well focused image region because wave front curve of sub-image region can be neglected. In experiment, the simulation result and real data processing result are shown. Well focused image is obtained by this algorithm.5. A novel frequency bandwidth combination algorithm is proposed for solving the discontinuity of phase. This method compensates range-independent slant range error. After that, range compression is implemented and frequency bandwidth is synthesized. The shared bandwidth is used to keep the phase continuous. Range migration correction is implemented. In the experiment, the simulation result and real data is processed. The range resolution of simulation and real data is analysized. The analysized result demonstrates the validity of the proposed method.6. A novel blind homomorphic deconvolution autofocus algorithm is proposed for enhancing the efficiency of motion compensation. The signal geometry is different from mulchannel autofocus algorithm. The similarity of these method is that the estimation of phase error is realized by sub-space smooth filtering. Ideology of the proposed method derives from cep-spectrum and homomorphic signal processing. As a non-linear filtering processing method, homomorphic signal filtering is proposed by A. V. Oppenheim in 1968 year. This technology has been applied to optical image, medical ultrasonic wave image, voice signal, communication signal reconstruction. In general case, blind homomorphic deconvolution can obtain well focused image without knowing the prior information about the defocused image. In medical field, 2 dimension point spread function(PSF) can be obtained by blind homomorphic deconvolution. It has given us enlightenment that is applied in medical field. A new method is proposed for estimating phase error of subaperture data. Further, the obtained subaperture phase error is combined into full aperture phase error. This method can be decomposed into several steps: 1) The introduction of geometry of homomorphic signal processing and blurring kernel; 2) The construction of smooth subspace; 3) The signal is filtered smoothly by the built Hibbert subspace; 4) The result of simulation and real data processing is given in experiment.7. Microwave staring imaging as a new radar imaging mode, depends on antenna sensor radiating random signal respectively and spatial-temporal two-dimension random radiation field is formed. Spatial-temporal two-dimension random radiation field vary with spatial position and time. The electromagnetic wave is radiated by antenna and collected by receiver. After that, the echo data and pre-stored random radiation field is processed by correlation algorithm. Then, targets in different resolution elements are distinguished. A novel staring imaging method is proposed in this part for suppressing high sidelobe by using Bjorck-Schmidt orthogonalization. The pre-stored 2 dimension random radiation field is orthogonalized by Bjorck-Schmidt method. The dependency of near units is reduced. The orthogonalized radiation field is processed combined with echo signal. Then, a well focused image is obtained in the experiment.8. Conclusion and expectation is summaried in final part.
|
PDF Full Text Request